AI has to earn its place.
AI belongs only where it solves a real problem better and its value remains transparent.
I am neither in favour of AI in every product nor fundamentally against it. Both reflexes avoid the same important question: What concrete problem does this solve better for a person?
The problem comes before the technology
Not every search needs a chatbot. Not every text field needs a suggestion, and not every process improves because a model is involved. When explicit rules solve a problem reliably, I prefer explicit rules. They are easier to test, explain and control.
AI becomes interesting to me where inputs are ambiguous, large amounts of information need structure or a first draft makes the next step easier. Even then, “with AI” is not a sufficient reason. The outcome must be faster, clearer or more useful than a simpler solution. Otherwise, the feature mainly creates additional cost, dependencies and failure modes.
somnara needs AI because language, personal context and recurring patterns do not fit into rigid rules. Webwoven, by contrast, does not need it in the game logic. Verifiable relationships and reproducible rules are the better solution there.
Value requires transparency
Anyone using AI takes responsibility for its limits. Users should be able to recognise when a result was generated, which data was processed and what they can do when it is wrong. Critical decisions need verifiable rules and a route to correction.
My standard is therefore not whether AI is technically possible or currently attracting attention. I ask whether I could explain its place in the product convincingly without the label. If the answer is no, it probably does not belong there. If the answer is yes, it still has to prove its value in everyday use, again and again.